Publication
Ludii as a Competition Platform
Authors:
Matthew Stephenson, Eric Piette, Dennis J. N. J. Soemers, Cameron Browne
Venue:
IEEE Conference on Games (CoG), 2019
Topics:
Ludii, competitions, general game AI, procedural content generation, human-like AI, game generation
Links: PDF · IEEE / ACM entry · arXiv
Abstract
Ludii is presented not only as a general game system for modelling, playing, and analysing traditional strategy games, but also as a platform for a wide range of AI competitions and challenges.
The paper compares Ludii with existing competition platforms such as GGP and GVGAI, and outlines future competition formats involving game-playing agents, procedural content generation, human-like AI, data mining, and other research directions.
It argues that the breadth, flexibility, and efficiency of Ludii make it a particularly strong basis for future general game competitions and for opening new avenues of AI research.
Context
This paper is important because it positions Ludii not just as a software platform, but as an infrastructure for future research communities and benchmarks.
It presents a broad vision for competition tracks that go beyond traditional general game playing, including procedural content generation, historical game reconstruction, explainable AI, deceptive games, human-like play, and mixed human-AI settings.
In that sense, the paper helps define Ludii as a long-term research ecosystem rather than a single tool, and shows how it could support many strands of future work in AI for games.
Full reference
Stephenson, M., Piette, E., Soemers, D. J. N. J., Browne, C. (2019). Ludii as a Competition Platform. In Proceedings of the IEEE Conference on Games (CoG). DOI: 10.1109/CIG.2019.8848084
BibTeX
@inproceedings{stephenson2019competition,
author = {Stephenson, Matthew and Piette, Eric and Soemers, Dennis J. N. J. and Browne, Cameron},
title = {Ludii as a Competition Platform},
booktitle = {Proceedings of the IEEE Conference on Games (CoG)},
year = {2019},
doi = {10.1109/CIG.2019.8848084},
url = {https://arxiv.org/abs/1907.00246}
}